monocular vision three-dimensional reconstruction method based on an improved Sift algorithm
A 3D reconstruction and monocular vision technology, applied in computing, image data processing, instruments, etc., can solve the problems of insufficient display of scene structure, shooting angle, precision constraints of feature extraction, unsatisfactory 3D reconstruction results, etc. Achieve good 3D reconstruction effect and improve matching accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0030] Example 1: See Figure 1-6 , a monocular vision three-dimensional reconstruction method based on the improved Sift algorithm, the specific process is as follows: step A, camera calibration. Study several calibration methods of the camera, and choose the most suitable calibration calculation method. Select a high-precision calibration plate and take images of the calibration plate from multiple angles. Detect the feature points in the image respectively, and then calculate the camera internal parameters and camera external parameters according to the selected calibration calculation method to complete the camera calibration;
[0031] Step B, image feature extraction and matching. Select appropriate image features, study image feature extraction methods and image feature description methods, obtain image target feature point sets, and then perform feature matching. Carry out software design, use Halcon image processing software, write programs, complete the realization...
Embodiment 2
[0034] Embodiment 2, on the basis of embodiment 1, improve the SIFT algorithm step as follows:
[0035] Step 1: Use a monocular camera to collect images from multiple angles, apply SIFT operator to any two target images and images to detect extreme points, and filter the obtained feature points to obtain the set sum of SIFT feature points;
[0036] Step 2: Calculate the phase consistency energy information matrix;
[0037] Step 3: Set the threshold value to a value between 0.01 and 0.05, compare it with the calculated phase consistency energy information matrix, and eliminate the feature point pairs whose maximum moment is lower than the threshold value, and obtain the set sum of feature point pairs to be matched;
[0038] Step 4: Perform stereo matching. Calculate the ratio of the closest point to the next closest point of the Euclidean distance and compare it with the set threshold to judge the similarity of the SIFT feature point pair. If the ratio is less than the set th...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


